Thread: Power

Wait, surely, Schildt must be an experienced programmer to write such books. He has written alot, too. On the cover of my book it says that "he is one of the most popular authors" and has sold "more than 2.5 million books".

But how then can he be wrong in so many cases and even contradictory to the standard? Aren't there some people that chack the book for errors before it's published?

Originally Posted by CornedBee

What Dae just said is wrong. There never was a cmath.h header in any standard. From the very beginning, the C++ standard contained math.h and cmath.

Its correct for some of the header files. I didnt know if cmath was one of those though (he said iostream). Regardless, they removed the .h and maybe changed the file name a tiny bit.

The CRT headers were "imported" into the C++ standard by removing the .h and prepending a c. The new C++ headers simply were created without a .h, as far as the standard is concerned. In practice, many of the C++ headers were headers that older compilers commonly used, with the .h removed.

Anyway, Dae, it doesn't change the fact that the standard was never changed. It wasn't TR1 (released in 2001) that changed the ending. So what you wrote was still wrong. And your "because" was also incorrect, because what you posted was an incorrect answer to the given question.

I'm not in a mood to be argued against, so you better be damn sure you've got your facts straight

Alright, better stop reading other peoples posts and only read yours then, because thats what I read. Sorry.

The CRT headers were "imported" into the C++ standard by removing the .h and prepending a c. The new C++ headers simply were created without a .h, as far as the standard is concerned. In practice, many of the C++ headers were headers that older compilers commonly used, with the .h removed.

So when <c___> showed up it was because they imported some other header, but made it so in practice it was similar to the use of <___.h>? If it wasnt already there already (which you said <cmath> has been around since the beginning) So what about <iostream>? it uses namespaces now, and has cout and stuff, so its an "import" too? when did they do that.. was it somewhat 4 years ago? or further back and people just didnt notice when they were making C++ books.

Before C++ had a standard, most compilers used <iostream.h>, <fstream.h>, <new.h>. etc for the C++ library headers. They also used <math.h>, <stdio.h>, <stdlib.h>, <time.h>, etc for the C library headers. The C library headers were already standard for the C language, so the same ones were used before C++ had a standard.

Then, in 1998, C++ got its first and only standard so far. In that standard, the headers used for the C++ libraries were <iostream>, <fstream>, <new>, etc. No mention was made of the old C++ headers, they were simply ignored and are non-standard. The C library headers were made part of the standard, so <math.h>, <stdlib.h>, etc are standard in C++. However, they were deprecated by the C++ standard. That means that you are allowed to use them, but it is recommended that you use the alternatives because some time in the future the deprecated versions might be removed from the standard. The new versions of the C library headers are <cmath>, <cstdio>, <cstdlib>, <ctime>, etc.

So, from the beginning of the C++ standard, the header files without the .h were the preferred ones. C++ was used before it was standardized, which is why many people are still confused about which headers to use.

In addition, the old C++ headers (like <fstream.h>) are different than the standard headers. You should not just mix and match. The deprecated C headers (like <math.h>) are the same as the new version (like <cmath>) except for the namespace changes. So in most cases, it doesn't matter in practice which C library header you use. If you have a choice, pick the new version.

Oh thanks Daved, that clears it up a lot.

So thats why.. because the C++ standard was made in 1998, tricky.